The application of next-generation sequencing (NGS) approaches to infectious disease diagnostics is anticipated to result in substantial clinical advances. In particular, techniques based on unbiased shotgun sequencing and targeted amplicon sequencing of DNA and RNA extracted from primary specimens are expected to have many advantages over classical microbiological approaches, including the detection of pathogens directly from primary specimens that may be difficult or impossible to culture. Nevertheless, routine application of this technique for diagnostic use is still challenging due to complicated requirements for sample preparation, the considerable expertise required for data analysis and interpretation, and the lack of off-the-shelf commercial solutions. This works aims to implement and validate an integrated NGS-based diagnostic system in a routine hospital microbiology service. Development of this system requires choosing and optimizing appropriate DNA extraction and library preparation methods for both ubiased and targeted amplicon sequencing approaches, assembly of a local computer and network infrastructure to support in-lab processing of data, implementation and validation of bioinformatics pipelines for data analysis, and clinical reporting. One exploratory aspect of this work is to determine the complexity of NGS pipeline output from clinical specimens, which has not been well characterized. The complexity of data obtained, in turn, will be used to guide the design of an interpretive framework for converting pipeline output into a form suitable for clinical reporting. In the current fiscal year, we have developed SOPs for nucleic acid extraction and NGS library preparation from primary clinical specimens, designed and implemented in-lab computer and network infrastructure resources capable of manipulating and analyzing NGS data, and installed and debugged software pipelines for NGS data analysis. Validation of the system for clinical use is currently in progress, and >100 samples representing various pathogen compositions and clinical specimen matrices have been sequenced and analyzed. In addition, we have estimated limits of detection using a metric that relates DNA concentrations, read counts and CFU counts. Initial validation results suggest that an NGS-based approach can offer superb sensitivity and specificity when variables are properly optimized.